CN-121686248-B - Construction method of rural housing structure safety multidimensional characterization index and evaluation system

CN121686248BCN 121686248 BCN121686248 BCN 121686248BCN-121686248-B

Abstract

The invention discloses a method for constructing a rural housing structure safety multidimensional characterization index and an evaluation system, and belongs to the technical field of safety identification, disaster prevention and reduction and artificial intelligence intersection of constructional engineering. The method comprises the steps of constructing a pesticide house structure safety quantitative characterization index system containing three layers of macro elements, micro elements and corresponding judging rules, collecting image data and geographical position information of a target pesticide house, carrying out collaborative analysis on the image data through an improved Mask R-CNN network and a panorama segmentation network based on a recursion layer aggregation structure, extracting quantitative parameter values strictly corresponding to the index system, inputting the extracted quantitative parameter values into a rule engine with the judging rules, automatically executing multi-level safety assessment by the rule engine, and outputting a final safety level and an assessment report. According to the invention, the efficient, objective and standardized agricultural house safety assessment is realized through the AI model and the rule engine, and the assessment efficiency and reliability are remarkably improved.

Inventors

LIU YIFENG
QIN PAN
XIE JUNQIAO
SONG QIANYI
JIANG XUE
MA XIAO

Assignees

中国建筑西南设计研究院有限公司

Dates

Publication Date: 20260512
Application Date: 20260210

Claims (8)

1. A method for constructing a rural housing structure safety multidimensional characterization index and an evaluation system is characterized by comprising the following steps: The method comprises the steps of 1, constructing a pesticide house structure safety quantitative characterization index system comprising three layers of macro elements, micro elements and a corresponding judging rule, wherein the macro element index is site dangerousness and is divided into three levels of dangers, potential dangers and safety according to geological disaster risks; Acquiring image data and geographical position information of a target agricultural house, carrying out collaborative analysis on the image data through an improved Mask R-CNN network and a panorama segmentation network based on a recursion layer aggregation structure, and extracting quantitative parameter values corresponding to indexes in an index system, wherein the collaborative analysis comprises the steps of associating and fusing a component instance list output by the improved Mask R-CNN network with a panorama segmentation map output by the panorama segmentation network, carrying out verification and refinement on the component instance by using pixel-level semantic information provided by the panorama segmentation map and establishing an accurate attribution relation between a damaged area and a component to which the damaged area belongs; And 3, inputting the extracted quantized parameter values into a rule engine with built-in judging rules, automatically executing multi-level security assessment by the rule engine, and outputting final security level and assessment report.
2. The method for constructing the multi-dimensional characterization index and the evaluation system for the safety of the rural housing structure according to claim 1, wherein the structural disaster prevention measure index in the mesoscopic element respectively prescribes the height and layer number limit value, the minimum wall thickness limit value and the earthquake resistant construction measure requirements for masonry, underframe, reinforced concrete, wood, stone, raw earth wall and steel structure.
3. The method for constructing the rural housing structure safety multidimensional characterization index and the evaluation system according to claim 2 is characterized in that the structure overall safety index in the microscopic elements comprises overall inclination rate and uneven settlement, the overall inclination rate is obtained by calculating the ratio of the horizontal offset value of the observation point to the height, and the uneven settlement is obtained by calculating the ratio of the vertical deformation difference of the observation point to the distance.
4. The method for constructing a rural housing structure safety multidimensional characterization index and assessment system according to claim 2, wherein in step 2, the input of the improved Mask R-CNN network is an image of a single Zhang Mubiao agricultural house, and the output is a bounding box, a pixel-level instance segmentation Mask, a component type label and a material type label of each detected structural component, so as to form a preliminary component instance list.
5. The method for constructing a rural housing structure safety multidimensional characterization index and evaluation system according to claim 4, wherein the improved Mask R-CNN network is specifically: Based on the original Faster R-CNN network, the ROI alignment operation is used to replace ROI Pooling; Based on the original Faster R-CNN network, adopting an improved ResNet and a characteristic pyramid network as a characteristic extraction trunk; a full convolution network is introduced in the split branches of the original fast R-CNN network.
6. The method for constructing the rural housing structure safety multidimensional characterization index and the evaluation system according to claim 5 is characterized in that the panoramic segmentation network is input into an image of a single Zhang Mubiao agricultural housing, and the image is output into a panoramic segmentation map with a semantic label and an instance ID for each pixel, wherein the semantic label comprises a plurality of structural member types, a plurality of damage types and a background type, and the damage types at least comprise cracks, flaking, weathering and rusting.
7. The method for constructing the rural housing structure safety multidimensional characterization index and the evaluation system according to claim 6, wherein the panoramic segmentation network comprises a feature extraction backbone network, a semantic segmentation head, an instance segmentation head and a panoramic fusion head.
8. The method for constructing a rural housing structure safety multidimensional characterization index and evaluation system according to claim 7, wherein the step 2 further comprises: and mapping the fused identified component and the damaged area to a three-dimensional space through the mapping relation between the two-dimensional image and the three-dimensional point cloud, and directly measuring the physical size of the component, the physical dimension of the damage and the overall inclination and uneven settlement parameters of the house.

Description

Construction method of rural housing structure safety multidimensional characterization index and evaluation system Technical Field The invention relates to the technical field of safety identification, disaster prevention and reduction and artificial intelligence intersection of building engineering, in particular to a construction method of a rural housing structure safety multidimensional characterization index and an evaluation system. Background The rural houses in China are huge in quantity and various in structure types, and a large number of houses are lack of normal design and construction in construction, so that potential safety hazards of different degrees are faced after the houses are used for a long time. At present, the safety identification of the rural houses is mainly based on the technical guidelines for rural housing safety identification. The method simplifies the flow, but the core of the method depends on manual field inspection, and qualitative descriptions such as slight cracking, serious deflection and the like are used as judgment basis, so that the evaluation result has strong subjectivity and poor consistency, and highly depends on personal experience of an identifier. Meanwhile, the manual multi-span inspection mode is low in efficiency and high in cost, and is difficult to realize the normalized safety general inspection of a large-scale agricultural house, and rapid safety screening and danger grading of massive damaged houses cannot be performed in emergency rescue of sudden disasters after earthquake and the like. In recent years, computer vision techniques based on deep learning have shown potential in the field of civil engineering health monitoring, for example for crack detection. However, the existing research is mostly limited to single and general lesion recognition, and cannot be deeply fused with a complete, professional and quantitative agricultural house safety assessment knowledge system. The fundamental problem is that accurate indexes (such as crack width limit value and wall thickness requirement of a specific type of component) meeting the definition of professional specifications cannot be automatically extracted from the image, and the extracted data cannot be embedded into a professional assessment flow comprising multi-level indexes, complex weight calculation and progressive decision logic. Therefore, developing a technical scheme capable of automatically and intelligently executing the professional assessment standard has become an urgent need for solving the problem of the safety supervision of the agricultural house. Aiming at the problems, a method for constructing a rural housing structure safety multidimensional characterization index and evaluation system is needed to solve the problems existing in the traditional method. Disclosure of Invention The invention aims to provide a rural housing structure safety multidimensional characterization index and evaluation system construction method, quantitative parameters are automatically obtained through an AI model, intelligent decision is made by utilizing a rule engine, high efficiency, objectivity and standardization of agricultural housing safety evaluation are realized, and evaluation efficiency and reliability are remarkably improved. In order to achieve the above purpose, the technical scheme adopted by the invention is as follows: a method for constructing a rural housing structure safety multidimensional characterization index and an evaluation system comprises the following steps: step 1, constructing a pesticide house structure safety quantitative characterization index system containing three layers of macro elements, micro elements and corresponding judging rules; Step 2, acquiring image data and geographical position information of a target agricultural house, and carrying out collaborative analysis on the image data through an improved Mask R-CNN network and a panoramic segmentation network based on a recursion layer aggregation structure to extract a quantization parameter value strictly corresponding to an index system; And 3, inputting the extracted quantized parameter values into a rule engine with built-in judging rules, automatically executing multi-level security assessment by the rule engine, and outputting final security level and assessment report. Further, the macro element index is site dangers and is divided into three levels of dangers, potential dangers and safety according to geological disaster risks, the micro element index comprises a structural disaster prevention measure index and a structural overall safety index, and the micro element index comprises quantitative description and grading thresholds aiming at cracks, deformation, section loss and node connection states of different material components. Further, the structural disaster prevention measure indexes in the microscopic elements respectively prescribe the height and layer number limit values, the minimum wa